Research on Knowledge Reduction based on Knowledge Granularity
Hong Li · Journal of Suzhou University · 2010
This paper,based on knowledge granularity and knowledge significance, proves that the granularity of the reduction attribute set is equal to the granularity of the initial attribute set,every attribute in the reduction set is important for each of the remaining attribute in the reduction set,every attribute out of the reduction set is unimportant for each of the remaining attribute in the reduction set,and every attribute in the core is important for each of the remaining attribute,etc.Then it describes the necessary and sufficient condition that core is a reduction.Finally,it offers knowledge reduction algorithm based on knowledge granularity-KRAKG algorithm,of which time complexity proves to be O(|R|×|U|2),effectiveness verified.